Abstract Retinoblastoma is a primary intraocular cancer that develops in the eyes of children. Tumorigenesis is initiated by a mutation in the RB1 gene, which was the first tumor suppressor gene described. Investigating the tumor suppressor pathway regulated by RB1 has provided unprecedented insights into the genetic mechanisms of tumorigenesis, not only for retinoblastoma, but virtually all human cancers. The evaluation of specific mutations, oncogenes and suppressors, tumor markers and gene expression profiles have revolutionized the care of breast, lung, prostate and other cancer patients. However, despite the fact that retinoblastoma-related research launched the field of cancer genetics, leveraging this knowledge to personalize the care for retinoblastoma patients has been elusive in large part because we cannot safely biopsy this tumor due to concerns of extraocular spread. Thus, we currently cannot correlate the genetic and genomic changes at the level of the tumor with clinical outcomes. Identifying these genomic changes, and providing clinical correlation is critically needed for retinoblastoma patients. The objective of this proposal is to develop a surrogate tumor biopsy for retinoblastoma using the aqueous humor, which is the clear fluid in a separate part of the eye from where the tumor forms. Starting in 2012, extractions of aqueous humor have been routinely and safely done as a standard part of the procedure to inject chemotherapy into the vitreous cavity of the eye; no cases of extraocular spread have been reported. Our preliminary data from evaluation of the aqueous revealed that tumor-derived cell-free DNA is present and whole genome sequencing demonstrates a profile of chromosomal gains and losses that corroborate changes in the tumor. The specific aims of this study are designed to test the following hypothesis: the aqueous can serve as a surrogate tumor biopsy ? more precisely, the aqueous harbors tumor-derived genetic material which can be assayed and correlated with clinical outcomes. Aim 1 is to define highly recurrent chromosomal changes in the aqueous, such as 2p gain/MYCN amplification, that are known to occur in retinoblastoma tumors. Aim 2 is to identify the specific RB1 mutation(s) that underlie tumorigenesis from the aqueous humor. Finally, Aim 3 will evaluate RNA expression profiles to define subgroups in the aqueous, and correlate these with clinical outcomes such as tumor recurrence. Currently, the only way any of these Aims are possible is from evaluation of tumor tissue, which is only available if an eye is removed (e.g. enucleated). However, by defining these genetic changes in the aqueous we will, for the first time, be able to characterize tumors in situ during therapy. This will allow for the development of prognostic markers for therapeutic response and, in the future, may guide therapy. Use of the aqueous humor as a surrogate tumor biopsy holds tremendous potential, from directly impacting the way we manage retinoblastoma patients, to ultimately allowing for the development of personalized medicine for this blinding childhood cancer. ! !